CDC revises decade-old foodborne illness estimates

Dec 15, 2010 (CIDRAP News) – The US Centers for Disease Control and Prevention (CDC) today released newer, more precise estimates of the foodborne illness burden, suggesting 1 in 6 Americans get sick each year, a rate lower than previously thought.

The new estimates for illnesses, hospitalizations, and deaths due to foodborne illnesses are the first since 1999, when a study that became known as "the Mead report" was published. The new findings from a CDC research group headed by Dr Elaine Scallan, assistant clinical professor at the University of Colorado at Boulder, appear today in two early online reports from the January 2011 issue of Emerging Infectious Diseases (EID).

Media and research reports have often cited the 1999 Mead report, which put the number of deaths from foodborne illnesses each year at 5,000. However, Scallan and her colleagues have lowered the mortality rate estimate to 3,000. The group also estimates that foodborne illnesses lead to 128,000 hospitalizations each year.

The CDC said today in background materials about the reports that the foodborne disease burden estimates can be used to set priorities for interventions, policy development, research and analysis, and advocacy. It added that regulatory agencies can use the estimates to conduct risk assessments during rulemaking.

Ramifications for food safety bill?
New findings come at a time when the Senate is poised, possibly this week, to vote on sweeping food safety legislation that is attached to a continuing resolution to fund the federal government for the 2011 fiscal year.

US Food and Drug Administration (FDA) Commissioner Dr Margaret Hamburg said in a CDC press release that foodborne illnesses and deaths are preventable and unacceptable. She said the nation can reduce the illness burden by intensifying efforts to implement prevention- and science-based measures.

"We are moving down this path as quickly as possible under current authorities but eagerly await passage of new food safely legislation that would provide us with new and long overdue tools to further modernize our food safety program," Hamburg said in the statement.

Why estimates are lower
The CDC said the new estimates are lower than the 1999 report, mainly because of improvements in the quality and quantity of data and new methods for estimating foodborne disease. The new estimates are the first to consider illnesses caused only by foods eaten in the United States.

The researchers separated their estimates into two parts, one on 31 major pathogens and one on unspecified agents, which include those without enough data to make estimates, not yet recognized as causing foodborne illness, and not yet discovered.

In its background materials about the new reports, the CDC said the new estimate's population survey sample size is five times larger than the one used in the 1999 report. The CDC said it used a stricter definition of gastroenteritis, which is says partly explains the lower illness rate.

Also, the researchers estimated that 26% of norovirus infections were foodborne rather than person-to-person spread, down from the 40% estimate made in 1999. The big drop in estimated foodborne norovirus infections resulted in significant lowering of the proportion of all known pathogens, which also reduced the proportion of unspecified illnesses thought to be foodborne.

The CDC also said newer estimates of people seeking medical care for foodborne illness symptoms is more accurate and that the 2011 estimate includes pathogen-specific multipliers adjusted for underreporting and underdiagnosis. It also said the researchers for the first time modeled uncertainty for the estimates. They include credible intervals for each number that indicate a 90% probability that the actual numbers fall within the estimate ranges.

Not an indication of lower foodborne disease
Though the burden estimate in the 2011 report is lower, CDC officials say it doesn't show that levels of foodborne illnesses have declined. At a press briefing today, Dr Christopher Braden, acting director of the CDC's division of foodborne, waterborne, and environmental diseases, said the estimates, because they vary so much in the quality and quantity of data and the calculation method, aren't comparable. "The 1999 estimate used the best data available at the time," he said.

However, the burden estimates serve as useful benchmarks for policy decisions." If we could decrease the level of foodborne illnesses by 1%, we'd keep about a half million people from getting sick from the foods they eat," Braden said.

Of 48 million foodborne illnesses that occur each year, 9.4 million are from the 31 known pathogens, and the remaining 38 million are from the unspecified agents, the Scallan reports note.

About 90% of illnesses, hospitalizations, and deaths were linked to seven pathogens: Salmonella, norovirus, Campylobacter, Toxoplasma, Escherichia coli O157, Listeria, and Clostridium perfringens. The leading cause of hospitalizations and deaths was Salmonella, at 35% and 28%, respectively.

Norovirus caused nearly 60% of the estimated illnesses, though it was responsible for a much smaller proportion of severe illness, the research group found.

More frequent estimates?
At the media briefing today, reporters asked if some measures, such as population survey findings, used in the estimates could be used to make more regular, perhaps yearly, foodborne illness estimates. Dr Patricia Griffin, chief of the CDC's enteric diseases epidemiology branch, told CIDRAP News that the CDC hasn't found that the surveys are helpful for identifying trends, because the questions on them often change. The geographic area covered by FoodNet, which conducts population surveys, has changed over time, with some site-to-site variation that has also made it more difficult to track patterns, she said.

Griffith said notable year-to-year fluctuations for some pathogens are possible and that the Scallan estimates incorporate uncertainty to account for such variability.

The CDC said its researchers will likely estimate the foodborne illness again in 5 to 10 years.

Experts weight in
In an editorial in the same issue of EID, Dr J. Glenn Morris, Jr., who directs the Emerging Pathogens Institute at the University of Florida in Gainesville, wrote that estimating the burden of foodborne disease is daunting, because only a small portion of illnesses are reported through official systems and there are a large number of pathogens that can be transmitted through food.

He wrote that over the past decade, the Mead estimates have driven efforts to make the nation's food supply safer, but concerns over the high degrees of uncertainty over some parameters cast doubt on the estimates and led to requests for the CDC to update the estimates.

Morris said the developers of new estimates, which he called "Sons of Mead," used the same basic approach but include two important improvements: the advanced treatment of statistical uncertainty and variability and the inclusion of many different data sources, models, and assumptions.

The best way to gauge if food is has gotten safer since the last estimates were made is to look at trends from FoodNet, a lab-based sentinel surveillance system, he said. He noted that FoodNet data showed an initial decline in 1995 after the US Department of Agriculture made some regulatory changes, some year-to-year variability in diseases, but no evidence of sustained improvement.

Estimates of burden for particular pathogens are important for understanding risks but are insufficient on their own, Morris wrote. To design targeted interventions, public health experts need better food attribution data that measure illnesses in food-pathogen combinations, he said, adding that the estimates don't address at-risk groups or the long-term side effects, such as end-stage renal disease, of some foodborne illnesses.

The Scallan group's reports, however, are "critical steps in the right direction" toward arming federal officials with better data and stronger analytic capacity, he wrote.

Michael T. Osterholm, PhD, MPH, director of the University of Minnesota's Center for Infectious Disease Research and Policy, publisher of CIDRAP News, credited the Scallan group for its work on the new estimates, but he said he worried that some may misinterpret the findings as a sharp drop in foodborne illnesses.

"The data related to specific pathogens is still very helpful," he said. For example, he said public health officials are pleased about the drop in E coli O157:H7 and Campylobacter illnesses, but are concerned about a rise in vibriosis and a lack of decline in the levels of Salmonella infection.

Scallan E, Hoekstra RM, Angula FJ, et al. Foodborne illness acquired in the United States—major pathogens. Emerg Infect Dis 2011 Jan;17(1) (early online publication) [Full text]

Scallan E, Griffin PM, AnguloFJ, et al. Foodborne illness acquired in the United States—unspecified agents. Emerg Infect Dis 2011 Jan;17(1) (early online publication) [Full text]

Morris JG. How safe is our food? (Editorial) Emerg Infect Dis 2011 Jan;17(1) (early online publication) [Full text]

See also:

Dec 15 CDC press release

CDC foodborne illness burden estimate background

Newsletter Sign-up

Get CIDRAP news and other free newsletters.

Sign up now»


Unrestricted financial support provided by

Bentson Foundation Unorthodox Philanthropy logo and text 'Leading Underwriter'3M logoGilead 
Grant support for ASP provided by


  Become an underwriter»